MIT Introduces New Power-Efficient Converter for Reducing Power Consumption in IoT Devices

MIT researchers have designed a power-efficient converter that could add a new asset to IoT device designers' toolbox.

Energy efficiency in the Internet of Things (IoT) devices has become a critical concern for designers due to decreasing device sizes and limited energy storage capacities.
The majority of IoT devices are usually deployed in hard-to-reach areas, or where replacing batteries is a challenge. As such, actuators and sensors should ideally run for years from small power sources such as AA or coin-cell batteries.
By design, most of these devices spend a good part of their time in the low power modes. They only activate when there is need to perform a particular task, after which they should quickly return to their low-power state. However, the challenge is designing circuits with the highest efficiency when in operation and yet very little, if any, leakage current when in the idle state.
Let's check out MIT's new efficient converter. But first, we'll take a high-level look at some of the issues of powering IoT devices and how designers are facing them.

Challenges in Powering IoT Devices and Other Small, Battery-Operated Devices

• Mobile devices and laptops use powerful operating systems to activate various power-saving modes such as entering into hibernation or dimming the display after periods of inactivity. However, the use of high-level operating systems is practically impossible for low-cost IoT sensors and actuators due to limited on-chip memory.

• A majority of power converters consume power even when in the idle mode or not providing any current to the load. The circuit may draw a microamp current or more to maintain this quiescent power. Sometimes, this might even be higher than what some low-power functions in the device requires.

• Maintaining efficiency over a wide range of currents—from deep sleep or standby to full device operation—is often not possible. For example, a device will draw a few microamperes when in the sleep mode but will require tens or hundreds of milliamps when active and performing energy-intensive activities such as data transmission. This presents a challenge to designers as they try to handle a ratio of 1,000,000:1 between the minimum and maximum current levels.

Energy Saving Strategies for IoT Devices

Achieving overall energy efficiency requires a variety of energy-saving strategies. Included in these strategies are selecting efficient hardware and design, as well as efficient software routines and codes. The converter design, hardware components, software code, data processing, encryption, and network protocols all impact on the overall energy requirements of the device.
It is essential to understand the device's energy profile both in active and idle modes. From here, designers can analyze the circuit operation, processes, and software routines to establish where most power is used or lost and if there are any energy saving opportunities.

The Waspmote Mote Runner dev platform for 6LoWPAN. Image courtesy of Libelium

Below are some of the different approaches that can help designers in reducing the dynamic and static energy consumption in low-power, battery-operated devices. These can be used in combination or individually depending on the specific needs of the application.

Reducing Leakage in Transistor Switches

Transistors in electronics switches are not completely perfect insulators and will allow small currents to flow, even when in off or idle states. Although the leakage current is usually very small, it can add up to a significant amount of power losses since many IoT devices spend most of their time in an idle state.
Applying a negative charge at the gate ensures that the device goes off completely and no current flows through when the transmitter is idle and the transistor is in the off-state. However, this is only effective if generating the negative charge requires much less energy compared to what the circuit would lose through leakage.

Reducing Power Consumption of the Transmitter Circuit

The transmission frequency is usually a function of the voltage. Most designs may therefore require high circuit voltage to take care of all the frequencies. However, by splitting the RF signal into discrete steps, a low-voltage converter can power most of the steps at the lower end and then boost the local voltage for the fewer steps at the upper spectrum.
Such a converter circuit uses inductors and capacitors to increase the low circuit voltage to the level required by the few steps that require higher values. With this approach, it is possible to design circuits with a low overall voltage and still have the ability to transmit higher frequencies efficiently.

Generating Several Voltage Levels from One Internal Power Source

Some devices utilize power management integrated circuits (PMICs) to provide several voltage levels from a single source. This makes it possible to drive various circuits with different voltage requirements and each of the rails can be tuned to provide just enough power for that particular application.
This enables using a low voltage to supply the majority of circuits and only generate high power for the specific applications that need it. PMICs may include functionalities like reset, watchdog timer, and more.

Texas TPS65290 ultra low power management IC

Selecting and Optimizing IoT Device Hardware

Proper selection of the microcontroller, which consumes a large portion of the energy in a device, is critical. The microcontroller needs to be energy efficient during operation and also have the ability to quickly wake up from sleep or any other low power mode.
But designers should also pay attention to all the other components, as well. The main energy consumers in IoT devices include sensors, wireless interface, voltage regulators, memory devices, and power management circuits.
While a good hardware and software design is crucial, the sensor’s circuit should distinguish between actual signals and ambient conditions such as noise in speech recognition devices. This prevents the IoT device from having to wake up and power all components due to false signals.
Below are the impacts on energy consumption by different components and applications based on the device type and features:

• An MCU without encryption hardware would require running software algorithms, hence going through multiple cycles that directly impact power consumption.
• Integrating fast, non-volatile, on-chip memory can yield significant power savings in IoT devices. While EEPROM devices require 6 mA and 3mA for a read and write operation respectively, energy efficient Ferroelectric RAM (FRAM) chips draw about 200 μA.The FRAM combines fast speeds with its non-volatile data storage, hence consuming much less energy both in active and sleep modes since it still retains information even when completely powered down.
• Integrated sensors with built-in signal conditioning are more energy efficient compared to the discrete devices which usually place additional processing requirements on the microcontroller. In addition, the discrete solutions require the MCU to remain active for longer periods.
• The choice of IoT protocol influences other functions such as the flexibility, power drain, reliability, range, and security—all of which have specific power needs.

Optimizing IoT Device Software

Minimizing the energy losses in the hardware is not enough. It's also necessary to optimize software processes and make them more energy efficient. This optimization includes determining and optimizing the software routines that seem to use the most energy. For example, designers can minimize the wake-up time and make the device remain in a low-power or sleep or mode for the longest time possible.

A New Solution: MIT's New Energy-Efficient Converter

The researchers from MIT Microsystem Technologies Laboratories are already working on an efficient converter design for the small power devices such as the IoT sensors, wireless radios, and other small electronic equipment.
This is a step-down converter that works with input voltages of between 1.2 and 3.3 volts, which it then reduces to 0.7 and 0.9 volts. Instead of supplying continuous energy like conventional converters, the MIT converter just sends enough energy packets that the circuit requires to perform a certain task.
The converter uses electronics switches, a capacitor, an inductor, and a variable clock that turns the switches on and off at different rates depending on the power requirements. When the device is in sleep mode, only a few packets are sent, but when there is need to perform a high-energy task, such taking a measurement and transmitting the data, the converter sends millions of energy packets, enough to accomplish the task.
The MIT energy efficient converter does the following:

• Reduces the amount of power a device uses when in operation by just supplying what the circuit requires
• Reduces the energy consumption when in sleep or rest mode by about 50 percent
• Maintains the power efficiency over a much wider range of currents spanning from 500 picoamps to 1 milliamp

This approach results in about 50 percent reduction in the quiescent or resting power requirements hence realizing huge energy savings.

---

Increasing the power efficiency in IoT and other low power, battery-operated devices requires a number of strategies that reduce energy consumption during both the active and sleep modes. Factors such as the converter design, hardware components, and software codes have an impact on the way a device uses energy.
Reducing the power consumption extends the battery life significantly, or enables the powering of small devices from ambient energy sources. MIT's new converter design could allow designers to develop a wide variety of low energy applications that would allow IoT devices to become truly efficient.


Featured image used courtesy of MIT